Microsoft Microsoft Certifications DP-100 Questions & Answers

• Question 381:

  You register a model in an Azure Machine Learning workspace by running the following code:

  

  You are creating a scoring script to use in a real-time service for the model.

  You need to write code in the scoring script to set the path of the registered model so that it can be loaded by the service. You include the necessary import statements.

  Which code segment should you use?

  A. path = Model.get_model_path(`loan_model')

  B. path = `model.pkl'

  C. path = ws.models(`loan_model')

  D. path = `outputs/model.pkl'

  Correct Answer: A

  Registering a model creates a logical container for the one or more files that make up your model. In addition to the content of the model file itself, a registered model also stores model metadata, including model description, tags, and

  framework information, that is useful when managing and deploying the model in your workspace.

  get_model_path

  Return the path to model.

  Note: The function will search for the model in the following locations.

  If version is None:

  Download from remote to cache (if workspace is provided)

  Load from cache azureml-models/$MODEL_NAME/$LATEST_VERSION/

  ./$MODEL_NAME

  If version is not None:

  Load from cache azureml-models/$MODEL_NAME/$SPECIFIED_VERSION/

  Download from remote to cache (if workspace is provided)

  static get_model_path(model_name, version=None, _workspace=None)

  Parameters

  model_name

  str

  Required

  The name of the model to retrieve.

  version

  int

  default value: None

  The version of the model to retrieve. Defaults to the latest version.

  _workspace

  Workspace

  default value: None

  The workspace to retrieve a model from. Can't use remotely. If not specified only local cache is searched.

  Returns

  The path on disk to the model.

  Reference:

  https://learn.microsoft.com/en-us/python/api/azureml-core/azureml.core.model.model

• Question 382:

  You are using a ScriptRunConfig object to configure an experiment that uses a script to train a machine learning model.

  The script must apply a regularization rate hyperparameter to the algorithm that is used to train the model.

  You need to pass the regularization rate in a variable named reg_rate to the script.

  Which code segment should you use?

  

  A. Option A

  B. Option B

  C. Option C

  D. Option D

  Correct Answer: B

  ScriptRunConfig Class

  Represents configuration information for submitting a training run in Azure Machine Learning.

  Constructor

  ScriptRunConfig(source_directory, script=None, arguments=None, run_config=None, _telemetry_values=None, compute_target=None, environment=None, distributed_job_config=None, resume_from=None,

  max_run_duration_seconds=2592000, command=None, docker_runtime_config=None)

  Parameters include:

  arguments

  list or str

  Required

  Optional command line arguments to pass to the training script. Arguments are passed in pairs, for example, ['璦rg1', arg1_val, '璦rg2', arg2_val].

  Example:

  # configure and submit your training run

  config = ScriptRunConfig(source_directory='.',

  script='train.py',

  arguments=['--arg1', arg1_val, '--arg2', arg2_val],

  compute_target=cluster,

  environment=env)

  Incorrect:

  Not A:

  Parameter _telemetry_values

  dict

  Required

  Internal use only.

  Reference:

  https://learn.microsoft.com/en-us/python/api/azureml-core/azureml.core.script_run_config.scriptrunconfig

• Question 383:

  You are implementing hyperparameter tuning for a model training from a notebook. The notebook is in an Azure Machine Learning workspace.

  You must configure a grid sampling method over the search space for the num_hidden_layers and batch_size hyperparameters.

  You need to identify the hyperparameters for the grid sampling.

  Which hyperparameter sampling approach should you use?

  A. uniform

  B. qlognormal

  C. choice

  D. normal

  Correct Answer: B

  Discrete hyperparameters (as this is Grid Sample, see note below)

  Discrete hyperparameters are specified as a Choice among discrete values. Choice can be:

  one or more comma-separated values

  a range object

  any arbitrary list object

  Python

  from azure.ai.ml.sweep import Choice

  command_job_for_sweep = command_job(

  batch_size=Choice(values=[16, 32, 64, 128]),

  number_of_hidden_layers=Choice(values=range(1,5)),

  )

  In this case, batch_size one of the values [16, 32, 64, 128] and number_of_hidden_layers takes one of the values [1, 2, 3, 4].

  The following advanced discrete hyperparameters can also be specified using a distribution:

  QUniform(min_value, max_value, q) - Returns a value like round(Uniform(min_value, max_value) / q) * q

  QLogUniform(min_value, max_value, q) - Returns a value like round(exp(Uniform(min_value, max_value)) / q) * q

  QNormal(mu, sigma, q) - Returns a value like round(Normal(mu, sigma) / q) * q

  *-> QLogNormal(mu, sigma, q) - Returns a value like round(exp(Normal(mu, sigma)) / q) * q

  Note: Grid sampling

  Grid sampling can only be employed when all hyperparameters are discrete, and is used to try every possible combination of parameters in the search space.

  For example, in the following code example, grid sampling is used to try every possible combination of discrete batch_size and learning_rate value:

  from azureml.train.hyperdrive import GridParameterSampling, choice

  param_space = {

  '--batch_size': choice(16, 32, 64),

  '--learning_rate': choice(0.01, 0.1, 1.0)

  }

  param_sampling = GridParameterSampling(param_space)

  Reference:

  https://learn.microsoft.com/en-us/azure/machine-learning/how-to-tune-hyperparameters

• Question 384:

  You create an Azure Machine Learning workspace. You are implementing hyperparameter tuning for a model training from a notebook.

  You must configure a Bandit termination policy that provides the following outcome:

  If the value of the primary metric of AUC is 0.8 at the point of evaluation intervals, any run with the primary metric value below 0.66 will be terminated.

  You need to identify which Bandit termination policy configuration to use.

  What should you identify?

  A. Set slack_amount to 0.2.

  B. Set slack_factor to 0.1.

  C. Set slack_factor to 0.2.

  D. Set slack_amount to 0.1.

  Correct Answer: C

  Bandit policy

  Bandit policy is based on slack factor/slack amount and evaluation interval. Bandit policy ends a job when the primary metric isn't within the specified slack factor/slack amount of the most successful job.

  Specify the following configuration parameters:

  slack_factor or slack_amount: the slack allowed with respect to the best performing training job. slack_factor specifies the allowable slack as a ratio. slack_amount specifies the allowable slack as an absolute amount, instead of a ratio.

  For example, consider a Bandit policy applied at interval 10. Assume that the best performing job at interval 10 reported a primary metric is 0.8 with a goal to maximize the primary metric. If the policy specifies a slack_factor of 0.2, any training

  jobs whose best metric at interval 10 is less than 0.66 (0.8/(1+slack_factor)) will be terminated.

  Reference: https://learn.microsoft.com/en-us/azure/machine-learning/how-to-tune-hyperparameters

• Question 385:

  You need to evaluate the potential risk of exposing personal information based on the values of epsilon and delta for differential privacy. You create a privacy report. What does an epsilon value greater than one represent?

  A. The privacy of data is preserved and there is limited impact on data accuracy.

  B. There is a high risk of exposing the actual data that is uses to generate the report.

  C. The data used in the report is very noisy.

  Correct Answer: B

  Epsilon: Put simplistically, epsilon is a non-negative value that provides an inverse measure of the amount of noise added to the data. A low epsilon results in a dataset with a greater level of privacy, while a high epsilon results in a dataset that is closer to the original data. Generally, you should use epsilon values between 0 and 1. Epsilon is correlated with another value named delta, that indicates the probability that a report generated by an analysis is not fully private.

  Reference: https://github.com/MicrosoftLearning/mslearn-dp100/blob/main/13%20-%20Explore%20Differential%20Privacy.ipynb

• Question 386:

  You create an Azure Machine Learning workspace named workspaces. You create a Python SDK v2 notebook to perform custom model training in workspaces.

  You need to run the notebook from Azure Machine Learning Studio in workspaces.

  What should you provision first?

  A. default storage account

  B. real-time endpoint

  C. Azure Machine Learning compute cluster

  D. Azure Machine Learning compute instance

  Correct Answer: C

  Reference: https://learn.microsoft.com/en-us/azure/machine-learning/v1/tutorial-1st-experiment-hello-world

• Question 387:

  You plan to run a script as an experiment. The script uses modules from the SciPy library and several Python packages that are not typically installed in a default conda environment.

  You plan to run the experiment on your local workstation for small datasets and scale out the experiment by running it on more powerful remote compute dusters for larger datasets.

  You need to ensure that the experiment runs successfully on local and remote compute with the least administrative effort.

  What should you do?

  A. Leave the environment unspecified for the experiment. Run the expenment by using the default environment.

  B. Create a config.yaml file that defines the required conda packages and save the file in the experiment folder.

  C. Create and register an environment that includes the required packages. Use this environment for all experiment jobs.

  D. Create a virtual machine (VM) by using the required Python configuration and attach the VM as a compute target. Use this compute target for all experiment runs.

  Correct Answer: C

  If you have an existing Conda environment on your local computer, then you can use the service to create an environment object. By using this strategy, you can reuse your local interactive environment on remote runs.

  Reference: https://docs.microsoft.com/en-us/azure/machine-learning/how-to-use-environments

• Question 388:

  You create an Azure Machine Learning workspace named workspace1. The workspace contains a Python SDK v2 notebook that uses MLflow to collect model training metrics and artifacts from your local computer.

  You must reuse the notebook to run on Azure Machine Learning compute instance in workspace1.

  You need to continue to log metrics and artifacts from your data science code.

  What should you do?

  A. Instantiate the job class.

  B. Instantiate the MLCIient class.

  C. Log in to workspace1.

  D. Configure the tracking URL.

  Correct Answer: D

  Track runs from your local machine or remote compute.

  Tracking using MLflow with Azure Machine Learning lets you store the logged metrics and artifacts runs that were executed on your local machine into your Azure Machine Learning workspace.

  Set up tracking environment

  To track a run that is not running on Azure Machine Learning compute (from now on referred to as "local compute"), you need to point your local compute to the Azure Machine Learning MLflow Tracking URI.

  Note:

  You can get the Azure Machine Learning MLflow tracking URI using the Azure Machine Learning SDK v1 for Python. Ensure you have the library azureml-sdk installed in the cluster you are using. The following sample gets the unique

  MLFLow tracking URI associated with your workspace. Then the method set_tracking_uri() points the MLflow tracking URI to that URI.

  1.

  Using the workspace configuration file:

  ws = Workspace.from_config() mlflow.set_tracking_uri(ws.get_mlflow_tracking_uri())

  2.

  Using the subscription ID, resource group name and workspace name:

  from azureml.core import Workspace

  import mlflow

  #Enter details of your Azure Machine Learning workspace

  subscription_id = ''

  resource_group = ''

  workspace_name = ''

  ws = Workspace.get(name=workspace_name,

  subscription_id=subscription_id,

  resource_group=resource_group)

  mlflow.set_tracking_uri(ws.get_mlflow_tracking_uri())

  Reference: https://learn.microsoft.com/en-us/azure/machine-learning/v1/how-to-use-mlflow

• Question 389:

  You use Azure Machine Learning to train a model.

  You must use Bayesian sampling to tune hyperparameters.

  You need to select a learning_rate parameter distribution.

  Which two distributions can you use? Each correct answer presents a complete solution.

  NOTE: Each correct selection is worth one point.

  A. Uniform

  B. Choice

  C. QNormal

  D. Normal

  E. LogUniform

  Correct Answer: AB

  Bayesian sampling only supports choice , uniform , and quniform distributions over the search space.

  from azureml.train.hyperdrive import BayesianParameterSampling from azureml.train.hyperdrive import uniform, choice param_sampling = BayesianParameterSampling( { "learning_rate": uniform(0.05, 0.1), "batch_size": choice(16, 32, 64, 128) } )

  Reference: https://learn.microsoft.com/en-us/azure/machine-learning/v1/how-to-tune-hyperparameters-v1

• Question 390:

  You create an Azure Machine learning workspace.

  You must use the Azure Machine Learning Python SDK v2 to define the search space for discrete hyperparameters. The hyperparameters must consist of a list of predetermined, comma-separated integer values.

  You need to import the class from the azure.ai.ml.sweep package used to create the list of values.

  Which class should you import?

  A. Choice

  B. Randint

  C. Uniform

  D. Normal

  Correct Answer: A

  Define the search space

  Tune hyperparameters by exploring the range of values defined for each hyperparameter.

  Hyperparameters can be discrete or continuous, and has a distribution of values described by a parameter expression.

  Discrete hyperparameters

  Discrete hyperparameters are specified as a Choice among discrete values. Choice can be:

  one or more comma-separated values

  a range object

  any arbitrary list object

  Example:

  from azure.ai.ml.sweep import Choice

  command_job_for_sweep = command_job(

  batch_size=Choice(values=[16, 32, 64, 128]),

  number_of_hidden_layers=Choice(values=range(1,5)),

  )

  Reference:

  https://learn.microsoft.com/en-us/azure/machine-learning/how-to-tune-hyperparameters